(CALMATTERS) – As artificial intelligence tools get better at mimicking human-created content, it’s soon likely that a college admissions essay written by a bot won’t read that much differently than one written by a human student.
It’s this concerning prospect, CalMatters’ education reporters Carolyn Jones and Mikhail Zinshteyn write, that is prompting some California colleges to be on alert. As part of its fraud policy, the college application tool Common App — used by the University of Southern California, Stanford University, Caltech and 1,000 other institutions nationwide — added a restriction in August on “substantive” use of AI in admissions applications.
The University of California operates its own admissions portal, and while it permits the use of AI in admissions essays, it says the practice should be limited — such as using it for “advice on content and editing.”
Even if colleges outright ban the use of generative AI to write personal statements, identifying plagiarism is difficult and imprecise. In March, a study found that humans can detect AI-written work about 50% of the time — basically, a coin flip. And even if an AI-content detection tool were developed with only a 1% false-positive rate, that would still mean 10 students for every 1,000 essays could be falsely accused of academic theft, wrote Wired.
Other universities, however, have no formal policy on AI use in admissions essays. In 20-plus years in admissions for the University of San Francisco, one associate provost couldn’t recall any applicant who was admitted with a strong essay but weak grades.
As for students themselves, one high school senior told CalMatters that relying on AI is detrimental to their own education.
- Toby Reed, from Fremont High in Oakland:“It’s bad enough stealing content, but with ChatGPT you’re not even stealing from a real person…. It’s your future. You can’t plagiarize in school. You can’t do it at work. People like AI because it’s quick and easy, but it’s not good.”